Mar
22
I've stumbled upon an interesting recording recently. I found a Bill Gates' talk from 1989 on reddit . Being interesting in history of computing as I am, I couldn't resist listening to this >90 minutes talk. Note, that 1989 is before famous Windows 3.0 (Windows 2.0 was in the market). MS worked with IBM on OS/2 back then.
It's been an though-provoking experience.
First Mr. Gates describes the computer industry in 70s, when Microsoft was starting their business. (Did you know that their first product - Altair Basic - was written with an emulator written by Gates and Allen on PDP?) Today we take it for granted, but in these days computers from different manufacturer were not compatible with each other. Each one came with different OS and different software. Thus IBM PC (and particularly - it's licencing model) was something revolutionary.
It's also interesting to hear where famous 640kB barrier came from. BTW, it took 6 years to go from 16k to 640k - I mean when 8-bit architecture was no longer enough. It took over 15 years to go from 640k to 4G, to the limits of 32-bit computers. What will be the next span?
Then Gates talks about how computer industry will look like in incoming years. Some of the predictions turned out to be true, some were embarrassingly wrong.
One statement was particularly interesting to me. Mr Gates said that Unix notion of pipes was good, but not enough for the age of GUIs. He proposes composing objects as a logical extension and wants to give users opportunity to use that in his incoming software. What actually happened, OLE and COM were so difficult that not only users but also developers found it very difficult to use. Only now with the Web, the users are getting back the possibility to mix different widgets made of snippets of different sites.
It's worth listening how perspectives on performance, GUIs or hardware design looked like 18 years ago. It's also funny to hear somebody explaining what is WYSIWYG, Internet or, well Google Earth. 
Some notable citations:
"With the exception of one project, there is less then 10 people involved in any particular product."
"It's not a matter of throwing a lot of programmers onto something. It's a matter of throwing right 10 people or less on individual projects."
"I don't think companies [like IBM] will ever learn how to write great software."
"We're having hard time convincing that a mouse was a good thing."
"There is this notion that over time there will be some much memory that people can be kind of slobs. That's not true. There will always be a premium for excellent software design."
There was also one about Unix being a chaos that will never succeed, but I cannot locate it now.
Have fun listening!